Chronic Subdural Hematoma

Abstract

Background

Chronic subdural hematoma (cSDH) is typically a disease that affects the elderly. Neurosurgical evacuation is generally indicated for hematomas that are wider than the thickness of the skull. The available guidelines do not address the common clinical issue of the proper management of antithrombotic drugs that the patient has been taking up to the time of diagnosis of the cSDH. Whether antithrombotic treatment should be stopped or continued depends on whether the concern about spontaneous or postoperative intracranial bleeding, and a presumably higher rate of progression or recurrence, with continued medication outweighs the concern about a possibly higher rate of thrombotic complications if it is stopped.

Methods

In this article, we review publications from January 2015 to October 2020 addressing the issue of the management of antithrombotics in patients with cSDH that were retrieved by a selective search in the Pubmed and EMBASE databases, and we present the findings of a cohort study of 395 patients who underwent surgery for cSDH consecutively between October 2014 and December 2019.

Results

The findings published in the literature are difficult to summarize concisely because of the heterogeneity of study designs. Among the seven studies in which a group of patients on antithrombotics was compared with a control group, four revealed significant differences with respect to the risk of thromboembolic complications depending on previous antithrombotic use and the duration of discontinuation, while three others did not. In our own cohort, discontinuation of antithrombotics (including both plasmatic and antiplatelet drugs) was associated with thrombotic complications in 9.1% of patients.

Conclusion

These findings imply that the management of antithrombotics should be dealt with critically on an individual basis. In patients with cSDH who are at elevated risk, an early restart of antithrombotic treatment or even an operation under continued antithrombotic therapy should be considered.

Chronic subdural hematoma (cSDH) is a spontaneous or post-traumatic serous fluid collection containing blood that is located between the dura mater and arachnoid and develops over a period of three or more weeks.

cme plus

This article has been certified by the North Rhine Academy for Continuing Medical Education. Participation in the CME certification program is possible only over the internet at cme.aerztebatt.de. The deadline for submission is 24 March 2023.

cSDH is most common in the elderly. Its incidence seems to be rising as the population ages (1), but it varies (1.72 to 20.6 per 100 000 persons/year [2, 3]) as a function of multiple factors, including sex (male predominance) and age distribution. For example, a Finnish study (1) revealed a marked effect of age on incidence (1), with an increase from 18.2/100 000 in the seventh decade to 52.1/100 000 in the eighth, 130.3/100 000 in the ninth, and 125/100 000 in persons aged 90 and older (study period: 2011–2015). The rising use of antithrombotics also seems to favor the development of cSDH (4, 5); according to a case-control study, both anticoagulants and platelet inhibitors increase the risk of cSDH (6). There is often a history of minimal head trauma, but this alone does not explain the chronic, progressive accumulation of fluid in the subdural space (7). Angiogenic stimuli may promote the generation of fragile vessels, which in turn give rise to microhemorrhages that make the hematoma larger (7). Inflammation also seems to play a role in the pathophysiology of cSDH (3, 7, 8).

Symptoms and diagnostic evaluation

Because it enlarges slowly, the hematoma is often well tolerated at first, with mild symptoms or none. Yet experience shows that cSDH is only rarely an incidental finding. Patients often present with nonspecific symptoms such as headache, unsteady gait, or cognitive impairment (9). There may also be focal neurological deficits such as hemiparesis or aphasia (9). Along with normal pressure hydrocephalus, cSDH is considered one of the few treatable causes of dementia (10). A clinical challenge of cSDH is to suspect the diagnosis and order the appropriate imaging studies of the head on the basis of nonspecific symptoms alone. cSDH is reliably demonstrated by non-contrast computed tomography (CT, the most common study) and by magnetic resonance imaging (MRI) (11, 12) (Figure 1, eFigure 1).

Figure 1.

An 89-year-old woman who fell at home.

a) Computerized tomography (CT) of the head one day after the fall reveals a thin acute subdural hematoma (aSDH) over the right cerebral hemisphere. Because of the lack of mass effect and the absence neurologic symptoms, surgery was initially not indicated.

b) Over the ensuing 5 weeks, the patient became increasingly disoriented and her level of wakefulness was impaired. A new CT revealed a chronic subdural hematoma (cSDH) that was mainly hypodense compared to the neighboring brain tissue. There was evident compression of the right lateral ventricle, indicating mass effect. The cSDH was successfully evacuated via burr hole trephination.

eFigure 1.

An 83-year-old woman was found lying on the floor with a right hemiparesis. There was no history of trauma in the weeks prior to the event.

a) CT revealed a chronic subdural hematoma (cSDH), hypodense in relation to brain tissue, with some more acute components and septations.

b–d): MRI on the same day revealed (b) marked hyperintensity of the hematoma compared to the cerebral cortex in the FLAIR sequence and (c) isointensity in the T1 sequence.

After the administration of contrast medium (d), there was contrast enhancement of the inner and outer membranes of the SDH, and of the septations. The differential diagnosis of this finding includes empyema. Further MRI sequences can usually distinguish a subdural empyema from a chronic subdural hematoma.

CT, computed tomography; MRI, magnetic resonance imaging.

Treatment and problems of antithrombotics

An asymptomatic, incidentally discovered hematoma that is no thicker than the width of the skull can usually be observed at first, but cSDH that is symptomatic when diagnosed usually requires surgical treatment (3, 12). Operative techniques range from bedside puncture procedures to extended burr hole trephination with or without the insertion of a drain (3, 12). cSDH is usually of liquid consistency and thus easily treated through a small operative approach, yet it tends to recur, with reported recurrence rates ranging from 10% to 20% (3). Embolization of the middle meningeal artery is a newer treatment option that can be performed in addition to, or instead of, evacuation of the hematoma; initial data on this method are promising, but the published case numbers are low (13). Although cSDH carries a better prognosis than other types of intracranial bleeding, patients with cSDH suffer a higher mortality for up to one year after diagnosis than persons in an age-adjusted cohort (14– 16). For example, the average survival time in a group of patients with cSDH whose average age was 80.6 years (range, 65–89) was 4.4 years, while the expected average survival time among persons of that age is 6 years (hazard ratio: 1.94) (16).

The optimal management of antithrombotic medication remains an unresolved issue (12): medications are often discontinued for fear of acute bleeding and an increased risk of recurrence. There is still no generally accepted scheme for managing antithrombotic drugs in patients with cSDH, who often suffer from multiple comorbidities; uncertainty arises from the trade-off of opposing risks (thrombotic complications versus cSDH recurrence or acute bleeding into a cSDH) (12, 17). In particular, there are hardly any robust data on the management of the direct oral anticoagulant (DOAC) drugs that have been introduced from 2010 onward. Thus, the question of antithrombotic medication in patients with cSDH still needs to be scientifically addressed.

Literature review

A selective search for relevant publications (January 1, 2015, to October 1, 2020) was carried out in the PubMed and EMBASE databases (for details, see eMethods, eFigure 2). The aim was to identify clinical studies (randomized controlled trials, cohort studies, case-control studies, and case reports with >10 patients) of patients with cSDH focusing on the management of antithrombotic medication, specifically including documentation of thromboembolic events during follow-up. The time period mentioned was chosen to be as up-to-date as possible in order to reflect current issues, including the use of DOACs. All DOACs that are now available were approved by the European Medicines Agency by 2015 (most recently edoxaban, in 2015). The primary endpoint was defined as the number of thromboembolic complications depending on the management of antithrombotics. Because of heterogeneity in the management of antithrombotic drugs across studies, no quantitative analysis was possible. The literature search was registered in the PROSPERO database (CRD42021215982) and carried out as specified in the PRISMA statement (18).

Literature analysis

A total of 856 abstracts (search on January 18, 2021: EMBASE: 692; PubMed: 164) that were published in the specified time period were examined. In 49 of these articles, antithrombotic medication and/or thrombotic complications were mentioned in the abstract; the full text of these 49 articles was analyzed, and 11 articles (19– 29) ultimately met the predefined inclusion criteria (efigure 2).

eFigure 2.

Flow chart of the literature review. 11 publications were included in the analysis.

TC, thrombotic complications.

The average number of patients per study was 448 (range, 150–817), and three studies with 140, 198, and 211 patients were restricted to patients taking antithrombotic drugs (24, 25, 27). All studies showed the well-known male predominance of 62–75%. The duration of follow-up ranged from 36 days to six months. All studies included surgically treated patients. Overall, the percentage of patients taking antithrombotic drugs (excluding data from the three studies [24, 25, 27] that were restricted to such patients) was 39.4 ± 17.3% (1636 patients on antithrombotics out of a total of 3824).

The management of antithrombotic drugs was highly heterogeneous. These drugs were temporarily discontinued in all studies (exception: only a few patients taking acetylsalicylic acid [ASA]), and they were restarted after a widely varying interval ranging from 2 to 90 days. When antithrombotic medication was restarted early, pausing the drug for an interval of 3 to 10 days was recommended (19– 21, 23, 24). Early restarting of antithrombotic drugs was not associated with complications. Surgery under ASA was also associated with a low risk profile (25). Longer interruptions of anticoagulation (30 days postoperatively) have also been advocated (27).

The overall rate of thrombotic events was 81 in 3824 cases (2.1% in all studies combined; range, 0–3.3%). The three studies that included only patients on antithrombotics (24, 25, 27) were excluded from this analysis as well. A total of 34 thrombotic events were detected in 549 patients, corresponding to a cumulative risk of 6.2%; the management of antithrombotics varied widely across studies.

An association between the occurrence of thrombotic events and the discontinuation of antithrombotic medication was significant in five studies and insignificant in six. After the analysis was restricted to studies comparing patients on antithrombotics to a control group, there were four studies that found significant differences in the risk of thrombotic complications depending on prior antithrombotic drug use or the duration of discontinuation (20, 22, 26, 28), and three that did not find any such significant differences (19, 23, 29). In one study, the number of thrombotic complications was zero (n = 178); thus, no increased risk was found in this study from temporarily discontinuing antithrombotics, although the study involved only 40 patients on antithrombotics (21). Prior use of antithrombotics was not found to increase the risk of recurrent cSDH in five studies (20– 23, 26); in one study, prior anticoagulation (but not antiplatelet agents) (29) was associated with an increased recurrence rate, as was antithrombotic therapy with multiple drugs in another study (19).

In two studies of ASA use (24, 25), the duration of interruption of ASA had no effect on the rate of thromboembolic complications. An early restart of ASA was not associated with a higher cSDH recurrence rate. The case numbers were low in these two studies (140 and 198 patients, respectively), particularly in view of the study of more than one treatment regimens (short versus long ASA interruption). In another study that included only patients on antithrombotics (n = 211) (27), longer pauses in antithrombotic medication were indeed associated with increases in thromboembolic complications. Moreover, in this study, the risk of recurrence was increased here if antithrombotic medication was restarted earlier than 30 days after surgery.

It should be pointed out that these studies were heterogeneous with respect to methodology, antithrombotic drugs (some studies dealt only with ASA monotherapy), only monotherapies with ASA were investigated), anticoagulant management (a pause after surgery of anywhere from 0 to 90 days’ duration), and methods of data analysis.

Adding up the case numbers of the six studies with data on anticoagulated as well as non-anticoagulated patients (19, 20, 21, 23, 26, 29), we find a 3.7% rate of thromboembolic complications in the former (33 of 883 patients) versus a 1.1% rate in the latter (18 of 1578 patients). This observed increased risk of thrombotic complications among patients on antithrombotics was statistically significant (Chi-square; p = 0.000014; odds ratio: 3.36; 95% confidence interval: [1,88; 6,01]), though the interpretability of this finding, derived from a simple addition of case numbers, is limited by the heterogeneity of the underlying studies. The findings of the individual studies are summarized in eTable 4. The authors of 9 of these 11 studies recommended restarting antithrombotic medication as early as possible (Box, Table, Figure 2).

eTable 4. Overview of studies.

Article	Year	Type of study	Case numbers	Age, sex	Intervention/ groups	Observation period	Anticoagulation	Management of anticoagulation	Thromboembolic complications (TC)	Outcome / Results	Conclusions
Zhang et al.	2020	retrospective, single-center	546 (January 2014– December 2017)	67 (58–78), male: 77 (62.10%)	burr hole trephination; groups: PAC n = 43 vs. PAI n = 81 vs. control group n = 422	at least 3 months	PAI n = 81 PAC (warfarin n = 28; DOACs n = 15) n = 43	pause, sometimes with reversal / delayed surgery (depending on neurological condition), restart 1 week after surgery	TC: all patients with AT: 7 (5.65%), PAC: 3 (6.98%), PAI: 4 (4.94%), control: 8 (1.90%) (p = 0.065)	AT: higher recurrence risk, mortality, length of stay; PAC: risk factor for recurrence (no difference between NOAC and warfarin)	AT: tendency to more severe course / worse neurological condition; PAC but not PAI is a risk factor for recurrence
Zanaty et al.	2020	retrospective, single-center	569 (2007–2015)	73 ± 13, male 66.67%	burr hole trephination	duration of post-operative follow-up	PAI: 46%PAC: 20% (among which > 98% warfarin)	subcategorization in: pause AT and restart after surgery 1) on day 2 2) on day 4–14 3) after 2–3 weeks 4) after 3–4 weeks 5) after 4–6 weeks	TC overall 0.9% (n = 6); risk factors for TC: univariate analysis: AT (OR: 6.030; p = 0.051); diabetes (OR: 4.835; p = 0.086); duration of AT pause (OR: 0.5; p = 0.182); multivariate analysis: AT (OR: 3.275; p = 0.049)	low risk of TC overall; AT restart ideally 2-21 d postoperatively, as there is a low risk of both stroke and hemorrhage in this period	optimal temporal window for restarting AT: 2–21 d postoperatively; machine learning model successfully predicts recurrent cSDH, but not TC (probably because of low case numbers)
Amano et al.	2020	retrospective, single-center	323 (2013–2018)	AT: 80 ± .,6, male 69%; control: 75.5 ± 11, male 62%	burr hole trephination; groups: AT (n = 108, of which n = 68 (63 %) PAI; n = 31 (29%) PAC; n = 9 (8%) AT + PAI) versus control (n = 215)	at least 24 weeks	PAI: n = 68 (63%); PAC: n = 31 (29%); PAC + PAI: n = 9 (8%)	PAC (of which n = 27 ‧warfarin, n = 13 DOAC): pause and preop. vit. K + PPSB i.v., restart in n = 37, new start (for DVD etc.) in n = 8 → n = 45 under AT postoperatively (in 80% start in the first two weeks after surgery); PAI: n = 77, of which in n = 21 OP after ≥ 1 week, after pausing and restarting in n = 17; n = 56 continuation of PAI without perioperative pause, n = 2 new start	TC in n = 8 (2.5%); no significant difference between patients with and without AT;TC in 4% (n = 2) of patients with continued PAI, 0% with pause of < 2 weeks’ duration, and in 43% (n = 3) of patients with pause of ≥ 2 weeks’ duration	hemorrhagic complication s n = 6 (1.9%) (no difference between patients with and without AT), TC without any significant difference	optimal restart of AT ‧ca. 3 days after surgery, multiple AT can elevate the recurrence rate
Poon et al.	2019	RCT (subsidiary analysis), multicenter (26 centers)	817, (May 2013– January 2014)	AT: 79 (73–84); male: 252 (71.4%).non-AT 75 (63–84); male: 301 (64.9%)	burr hole trephination	3 months	AT: 353 (43.2%); of which PAI (n = 171): ASA: 41.4%, clopidogrel: 6.8.%, double PAI: 4%AT (n = 148): warfarin 41.9%. double PAI +AT: 3.1%, no DOACs	Discontinuation 1–7 (AT) or 1–3 d (PAI) before OP; PAI: n = 58 (29.3%) platelets; PAC: vitamin K + PPSB (52.7%) PPSB (12.2%); vitamin K (23%)	TC: overall n = 15 (during hospitalization): n = 11/336 (3.3%) in the AT group and n = 4/449 (0.9%) in the group without AT more TC in the AT group (OR: 3.70; 95% CI [1.16; 11.80]; p = 0.02) compared to the group without AT, there was a higher risk of TC in the PAC group (OR: 4.02; 95 CI [.,06; 15.30]; p = 0.03), but not in the PAI group (OR: 2.75; 95% CI [0.68; 11.2]; p = 0.14)	recurrence risk: no significant difference between the AT and non-AT groups (10.1% vs. 9.9%)predictors of recurrence: male sex, low GCS score preoperatively, and no drainage. no difference in recurrence rate depending on whether PAI are paused 3 or 6 days preoperatively. significantly more TC in the AT group.	During hospitalization, n = 15 TC; significantly higher rate in the AT group; AT group: higher risk of TC but not of recurrent cSDH or worse outcome the findings do not support delayed OP in patients with AT
Dziedzic et al.	2017	retrospective, single-center	178 (2007–2011)	69.4 (20–99)male: 68.6% (132)	burr hole and/or subdural drainage / craniotomy with large septations (13 cases)	average, 3 months (1 - 6 months)	PAI: n = 5; 2.8.%); PAC: n = 35 (19.7%) No AT: 138; 77.5%)	pause preoperatively, if PAC reverse with FFP or PPSB, if PAI: in a stable patient, pause 7 days before surgery; in case of poor neurological status, platelets and then OP, one week after surgery restart AT (in case of PAI, restart ASS first)	No TC in the period of observation	no statistically significant difference in operations for recurrence in the groups with and without AT (AT 12 [9%], non-AT 3 [7.5%]). All operations for recurrence in the AT group were in patients who had been taking oral anticoagulants.	reversal of AT: no elevated risk of TC no TC with AT restart at 1 week; similar treatment outcomes in the two groups. AT did not elevate the recurrence rate.
Amano et al.	2016	retrospective, single-center	150 (2011–2015)	77 (30–102) male: 106 (70.7%)	burr hole trephination + drainage	at least 3 months	AT: n = 44 (29%) PAI: n = 22 (14.7%); PAC: n = 17 (11.3%), of which n = 1 DOAC; PAI + PAC n = 5 (3.3%)	PAC: pause and vitamin K + PBSB, PAI: pause without platelets (n = 20), no pause (n = 7). Restart AT (if no complications arose) in one week	TC in n = 5 (n = 4 with prior AT); incidence of TC 9.1% with prior AT versus 0.9 % without	same risk of postoperative hemorrhages and recurrences (AT versus non-AT); TC in n = 5 (elevated risk with prior AT)	AT did not affect the rate of hemorrhagic complications, but TC were more common in the AT group; AT should be restarted as early as possible after surgery (in the absence of hemorrhagic complications)
Guha et al.	2016	retrospective, single-center	479 (2000–2012)	72.3 ± 13.6 male: 341 (71.2%)	burr hole trephination, craniotomy, craniectomy, drainage	3.1 months	AT n = 231 (48.2%) PAI n = 170; PAC n = 186; DOAC n = 1, LMWH n = 8	AT pause an average 1 day preoperatively (clopidogrel: 3 days)PAC: n = 59 (69%) reversal with vitamin K and PPSB or FFP restart after an average of 52 days, 75% of patients had an AT pause for > 2 weeks. Restart of AT in only n = 120 (51.9%) patients overall	overall TC: n = 8 (1.67%), of which n = 6 with AT	TC: no difference between the preoperative AT and non-AT groups. (2.60% vs. 0.81%; OR: 3.28; 95% CI: [0.66; 5.90]; p = 0.16); TC arose significantly earlier in patients with prior AT (median 2.7 vs. 51.5 days; p = 0.036) Postoperative restarting of AT was associated with a lower rate of rebleeding (26.9% versus 2.2%; OR: 0.06; 95% CI: [0.02; 0.2]; p < 0.01)	restarting AT did not elevate the risk of recurrent SDH or of hemorrhagic complications. Restarting AT 3 days after surgery can be safe.
Fornebo et al.	2017	retrospective, single-center	763 (2010–2015)	no AT: 71 ± 13; AT: 78 ± 9;male: no AT: 65.1%, AT: 71.8%	burr hole trephination and drainage	4 to 8 weeks postoperatively 6 months for TC 3 months for mortality	no AT n = 455 (59.6%) AT n = 308, (40.4%). early vs. late restart	2 groups: early restart of AT (< 30 d postoperatively), n = 100/302 (33.1%). (on average: 15 d) late restart (> 30 d) n = 202/302 (66.9%), (on average 66 d)	TC were significantly more common in the late-restart group than in the early-restart group (11% vs. 2%)	1. reoperation: no difference between early and late restart 2. mortality by 3 months after surgery: no difference between early and late restart 3. significantly lower morbidity in the non-AT group	AT: no significant effect on recurrence rate or perioperative mortality; restarting AT is associated with a lower risk of TC
Studies that only included anticoagulated patients:
Kamenova et al.	2017	retrospective, single-center, subgroup analysis	198 (from an overall cohort of 963) (2009–2015)	77.5 ± 8.5 male n = 139 (70.2%)	burr hole trephination with drainage off suction for 48 hours	at least 36 days	only PAI with ASA, n = 198 2 groups continued ASA group n = 26 (13.1 %) (brief peri-interventional pause) n = 172 (86.9%)	ASA group: restart after 2.81 days (SD ± 2.56); ASA-pause group: restart after 51.24 days (SD ± 33.83)	TC: n = 14: n = 3 (11.5 %) in the ASA Group (only a brief pause), n = 11 (6.4%) in the control group (a longer pause)(p = 0.4); no significant difference with respect to myocardial infarction and other cardio-vascular/thromboembolic complications	no significant difference with respect to either recurrence rates or TC	recurrence rate not affected by ASA; acute bleeding equally common in both groups; TC equally common in both groups, surgery under ASA seems justifiable
Kamenova et al.	2016	retrospective, multicenter, subgroup analysis	140 with ASA (from an overall cohort of 750) (January 2009– June 2014)	77.3 ± 8.3male n = 106 (73.7.%)	burr hole trephination with drainage off suction for 48 hours	on average, 43.5 d postoperatively	only PAI with ASA n = 32: elective surgery, ASA pause for 5 days n = 108, with deficits: urgent OP: n = 80 (74.1%) ASA taken up to and including the day of surgery; n = 28 (25.95) 1– 4 day pause	individual decision about restarting ASA n = 32 (22.9%) ASA permanently discontinued postoperatively n = 108 restarted postoperatively (in the first 90 days after surgery)	overall: TC in 10 patients (7.1%), no significant difference between pausing and not pausing ASA	similar recurrence rates with early and late restarting of ASA; poor GCS score and mRS postoperatively were significant risk factors for recurrent SDH; TC in 10 of 140 patients (7.1%) (no effect of ASA intake)	comparable recurrence rates with early or late restarting of low-dose ASA therapy; restarting ASA after 7 to 10 days seems justifiable
Todeschi et al.	2020	prospective, multicenter, observational study	211 under AT (May 2014– March 2018)	79.6 ± 9.2; male: 158 (74.9%)	evacuation of cSDH (various techniques)	3 months	PAC: 49.3% (n = 104); NOAC: 1.,4% (n = 22); PAI: 62.1% (n = 131)	PAC: “stop and wait” n = 23 (221%), vitamin K alone n = 27 (26%), PPSB n = 17 (16%), PPSB + vitamin K n = 37 (35,6 %); PAI: no intervention/pause n = 43 (32.9%), “stop and wait” n = 55 (42%), platelets n = 32 (24.4%), FFP n = 1 (0.7%)	TC in n = 10 (4.7%), TC and recurrent cSDH n = 6 (2.8%), long-term pausing of AT was a significant risk factor for TC	not restarting AT elevates the OR for TC [OR: 7.5; 95% CI: [1.2; 42]; p < 0.00001]; higher recurrence rate when at are restarted in less than 30 days after surgery	TC in 8% of the overall patient collective, restarting AT after 30 days is recommended

AT, antithrombotic therapy; ASA, acetylsalicylic acid; CI, confidence interval; cSDH, chronic subdural hematoma; DOAC, direct oral anticoagulants; DVT, deep venous thrombosis; FFP, fresh frozen plasma; LMWH, low-molecular-weight heparin; mRS, modified Rankin score; NOAC, newer oral anticoagulant drug(s); OP, operation; OR, odds ratio; PAC, plasmatic anticoagulation (=DOAC or newer oral anticoagulants); PAI, platelet aggregation inhibition; PPSB, prothrombin complex concentrate; RCT, randomized, controlled trial; TC, thromboembolic complications

BOX. Clinical data.

In a dataset from a single center (for details on methods and results, see eMethods), data from 395 patients with chronic subdural hematoma (cSDH) were analyzed with regard to tthromboembolic complications (TC) arising within six months of diagnosis. Baseline data of the patient collective and subgroups, depending on the occurrence of TC, are shown in the Table. 20 patients (5.1%) sustained a thrombotic complication at some time in their clinical course: these complications included one myocardial infarction (0.3%), four pulmonary embolisms (1%), 10 strokes (2.5%), three deep venous thromboses (0.8%), and two other thromboembolic complications (0.5%). Mortality in the overall collective was 1.3% (5/395). One patient may have died of an underlying cancer; thus, thrombosis-related mortality was 1.0% of the overall collective (or 2% of the patients on antithrombitics). Of the 20 patients who sustained thrombotic complications, 18 (90%) were on antithrombotics before surgery (8 with phenprocoumon, 6 with ASA, 2 with a DOAC, 1 with clopidogrel, and 1 with ASA plus another substance). In the univariate model, the following factors were associated with the occurrence of thrombotic complications: pre-existing neurological disease, pre-existing pulmonary disease, pre-existing malignancy, pre-existing cardiac disease/arrhythmia, preoperative paresis, Markwalder score (dichotomized), and the use of antithrobitic drugs. In the multivariable model, the use of antithrombotic drugs before the intervention and their interruption up to the time of the intervention or beyond it (until recovery) was a highly significant predictor of thromboembolic complications (odds ratio [OR]: 8.28; 95% confidence interval: [1.8; 38.12]; p = 0.007). Thus, the likelihood of a thrombotic complication was more than eight times higher in cSDH patients who were previously treated with antinthrombotics than in those who weren’t. The individual cases of thrombotic complications (depending on whether or not the patients had been treated with antithrombotics before they were diagnosed with cSDH) are shown in Figure 2.

Table. Distribution of variables (stated as total numbers and percentages, with means and standard deviations) for the overall patient collective, with and without thromboembolic complications (TC).

	Overall (n = 395)	No TC (n = 375)	TC (n = 20)	p-value*
Age	75 ± 12	75 ± 12	79 ± 6	0.016
Sex (female)	145 (36.7%)	139 (37.1%)	6 (30%)	0.522
Markwalder score	1.6 ± 0.7	1.6 ± 0.7	2.0 ± 0.7	0.018
CHA2DS2-VASc score	3.4 ± 1.9	3.4 ± 1.9	3.9 ± 1.9	0.194
GCS score	14.2 ± 2.1	14 ± 2	13.7 ± 2.1	0.004
Antithrombotic medication	198	180 (48%)	18 (90%)	0.0002
Preoperative paresis	210 (53.7%)	195 (52%)	15 (75%)	0.075
Cardiac arrhythmias	95 (24.1%)	82 (21.9%)	13 (65%)	< 0.0001
Prior heart disease	138 (34.9%)	125 (33.3%)	13 (65%)	0.008
Prior lung disease	37 (9.4%)	33 (8.8%)	4 (20%)	0.200
Prior malignant disease	60 (15.2%)	54 (14.4%)	6 (30%)	0.115
Prior neurological disease	75 (19.1%)	74 (19.7%)	1 (5%)	0.179
Alcohol abuse	19 (4.8%)	18 (4.8%)	1 (5%)	0.620
Drug abuse	8 (2%)	8 (2.1%)	0	–

*p-values for comparisons of the groups with and without thrombotic complications (t test, chi-squared test, Mann-Whitney U test); significant findings with respect to the primary endpoint are in boldface.

GCS, Glasgow Coma Scale.

Figure 2.

The temporal course of thrombotic complications (TC) as a function of the use of antithrombotic drugs (AT).

Evaluation and perspectives

Chronic subdural hematoma is a common disease with relevance in everyday life. Its incidence has risen in recent years, possibly because multimorbidity has become more common, along with the rising use of antithrombotic drugs and the increased availability of neuroimaging (1). Most patients have a positive past medical history. There are not yet any guidelines for the optimal management of anticoagulation. Current data suggest that interrupting antithrombotic medication for longer times leads to a clinically significant rise in thrombotic events (9.1% among our own patients who were taking anticoagulants preoperatively), and that the early restarting of antithrombotics (depending on the indication) most likely does not lead to additional complications (literature review). The management of antithrombotics in cSDH patients must therefore be reevaluated, as current knowledge suggests that the early restarting of antithrombotic medication ought to be considered.

For a subset of these patients, especially those with an absolute medical indication for antithrombotic therapy with multiple drugs, embolization of the middle meningeal artery (MMA) may be a therapeutic alternative (30, 31).

The high rate of thrombotic events in our own cohort is most likely attributable to our restrictive use of antithrombotic drugs, with temporary discontinuation as our internal hospital standard. These drugs were managed more liberally, with a shorter interruption of antithrombotic medication, in the studies that were included in our literature review.

A large study based on the Medicare database (32) confirmed the clinical importance of cSDH and of the management of antithrombotic drugs in patients with this disease. Out of 1.7 million cases analyzed, there were 2939 with an atraumatic subdural hematoma. The affected patients were more likely to have pre-existing illnesses than the age-matched population without subdural hematoma, and they had a significantly increased risk of arterial ischemic complications (hazard ratio [HR]: 3.6; [1.9; 5.5]), and stroke in particular (HR: 4.2; [2.1; 7.3]), in the first 4 weeks after treatment for SDH, but not at later times. Subgroup analysis revealed a higher risk of arterial ischemic events among patients who had an indication for antithrombotic treatment, and the authors inferred that this may have been caused by intentional interruption of antithrombotic therapy for four weeks (in analogy to the American Heart Association recommendations for intracerebral hemorrhage [33]). In Germany, there are not yet any specific guidelines for antithrombotic management in patients with cSDH. AWMF guidelines on intracerebral hemorrhage are now in preparation.

Among the studies we analyzed, the one with the largest number of cases (a subsidiary analysis of patients included in a randomized controlled trial; n = 817 [26]), displayed a problem that was also prominent in the literature analysis as a whole: the management of anticoagulation in patients with chronic subdural hematoma is highly heterogeneous. It is based not only on individual, patient-specific factors, but also by historical circumstances (“local hospital tradition”).

In a systematic literature review and meta-analysis concerning anticoagulation in chronic subdural hematoma dating from 2014, only three studies were found suitable for inclusion, with a total of 64 patients (34). The authors concluded that there was little empirical evidence to support any particular mode of managing antithrombotics. A recent paper on the resumption of antithrombotic therapy after surgery for cSDH, summarizing data from eight primary publications, revealed no difference in the rate of hemorrhagic complications after the restarting vs. continued pausing of antithrombotic drugs (35); patients whose anticoagulants were restarted had significantly fewer thromboembolic complications (2.9% vs. 6.8%, p < 0.001).

A survey revealed uncertainty among physicians about whether they should restart anticoagulant drugs; widely varying approaches were found, with different estimations of the risk of thromboembolic complications on the one hand, and hemorrhagic complications on the other (36). The interpretation of the findings reported in this review is limited by the fact that all antithrombotic drugs were grouped together, and that thrombotic complications were defined to include as both arterial and venous vaso-occlusive events. For more detailed analyses, studies with higher case numbers will be needed. Any conclusions drawn from this literature review are tentative in view of the heterogeneity of the underlying studies in their methods and in the questions asked.

Overview

The incidence of chronic subdural hematoma is rising, in parallel with the demographic trend. There are, as yet, no guidelines for the use of antithrombotics in the treatment of this disease.

Conclusion

The data presented here tend to support the early restarting of antithrombotic drugs after the surgical treatment of cSDH, or even, in some cases, treatment under uninterrupted antithrombotic therapy in patients at particular risk. This represents a paradigm shift in the management of cSDH for quite a few patients. Early restarting of antithrombotic drugs seems to carry little risk, particularly as concerns recurrent cSDH. Continuing antiplatelet therapy with ASA is justifiable in view of the available data.

Supplementary Material

eMethods

Clinical data—Methods

Patients who had undergone surgery for cSDH at the Department of Neurosurgery, RWTH Aachen University Hospital, were identified and analyzed pseudonymously via an OPS query (OPS code 5–013.1, 10/01/2014 to 12/31/2019). Ethical approval was obtained for the collection and analysis of data (Ethics Committee of RWTH Aachen University; EK 399–20 of 1 November 2020), and the study was registered in the German Register of Clinical Trials (DRKS00025280). The findings are reported as recommended in the STROBE statement guidelines (e1).

Patients with a neurologic deficit (e.g., impaired consciousness or a focal deficit) underwent surgery on the day of hospital admission (hematoma evacuation via extended burr hole trephination, with drain insertion). For patients on antithrombotics (i.e., those taking plasma, antiplatelet, or combined anticoagulant drugs), reversal of the antithrombotic effect was attempted: vitamin K antagonists were antagonized with prothrombin complex concentrate [PPSB] and vitamin K substitution, antiplatelet drugs with platelet concentrate. Some patients taking acetylsalicylic acid (ASA) only were not given platelet concentrate, at the discretion of their neurosurgeon. Direct oral anticoagulant drugs (DOACs) were not antagonized. Antithrombotic medication was paused postoperatively until normalization of the imaging findings. The following data were documented: age, sex, past medical history, preoperative anticoagulation, neurologic status, drug/alcohol history, Markwalder score (prognostic score for cSDH; tabulated in eTable 1 [e2]), CHA₂DS₂-VASc score, and thrombotic complications (TC), which were defined to include all vaso-occlusive events, whether venous or arterial – as in Ref. [26] – due to thrombosis in the six months after surgery. Data on TC were collected from the patients and their primary care physicians by telephone. The factors affecting thrombotic complications were then studied by bivariate analysis (chi-squared test). For this purpose, the Glasgow Coma Scale (GCS), Markwalder, and CHA₂DS₂-VASc scores were dichotomized (GCS: 3–8 versus ≥ 9; Markwalder score: 0–1 versus 2–4; CHA₂DS₂-VASc score: 0–4 vs. 5–9; see [e3]). Parameters that were associated with thrombotic complications with a p-value below 0.2 were included in the multivariable model, as were age and sex as underlying covariates (binary logistic regression, method: inclusion). Multicollinearity could not be demonstrated, and the studentized residuals, leverage values, and Cook’s distances were checked. All of these are used, simply stated, to detect outliers. Studentized residuals are residuals divided by the estimated standard deviation, and are thus a measure of variability and of outliers. Leverage values reflect the influence of a value on the goodness of fit of a model. Cook’s distance is a measure of the influence of a single variable on the entire model (etable 1).

All values are given as means with standard deviations. The t-test, chi-squared test, and Mann-Whitney U test were used to make comparisons depending on the data structure. The analysis was performed with the aid of SPSS 25.0 (SPSS Statistics for Windows, Chicago, USA) and GraphPad Prism 9.0.2 (GraphPad Software, La Jolla, California USA).

Clinical data—Results

395 patients (age: 75 ± 12 years; 145 female, 250 male [36.7%, resp. 63.3%]) with cSDH were included (database query: 787 patients, of whom 392 were excluded because of acute hematoma [n = 178], subdural empyema [n = 2], age <18 years [n = 7], missing data [n = 42], missing contact information [n = 21], patient refusal [n = 53], other reasons [n = 89]). Antithrombotic drugs were taken by 198 patients, or 50.1% of the total collective: 87 (22%) were taking ASA alone, 4 (1%) clopidogrel alone, 23 (6%) ASS and another drug (which was a P2Y12 inhibitor for 9 patients, phenprocoumon for 6, a DOAC for 3, and a different drug for 5), 62 (16%) phenprocoumon, 19 (5%) a DOAC, 2 (0.5%) therapeutic heparinization, and 1 (0.3%) a different combination. In 31 cases, anticoagulation was reversed because of clinical urgency (with PPSB (± vitamin K) in 26 cases, platelet concentrate in 3 cases, and prothrombin in 2 cases). Antithrombotic medication was paused for a mean of 102 days (but with high variance: standard deviation [SD]: 98 days; range, 1 – 468 days). Baseline data for the total collective and for patients who were or were not taking antithrombotics preoperatively are shown in eTable 2 (as a function of preoperative use of antithrombotic medication prior to diagnosis) and in the Table (as a function of the occurrence of thrombotic complications) (etable 2).

20 patients sustained thrombotic complications (5.1%), including one myocardial infarction (0.3%), four pulmonary embolisms (1.0%), 10 strokes (2.5%), three deep vein thromboses (0.8%), and two other TCs (0.5%). [Note: these percentages were stated incorrectly by a factor of 100 in the German original.] Mortality was 1.3% of the overall collective (5/395). One patient may have died of an underlying cancer; thus, thrombosis-related mortality was 1.0% of the overall collective (or 2% of the patients on antithrombotics). Of the 20 patients who sustained thrombotic complications, 18 (90%) were on antithrombotics before surgery (8 with phenprocoumon, 6 with ASA, 2 with a DOAC, 1 with clopidogrel, and 1 with ASA plus another substance). The administration of thromboprophylaxis during the pause of antithrombotic therapy was documented in 176 (89%) of the patients who were on antithrombotics before surgery. Sixteen acute rebleeds were noted (but often without clinical relevance) 12 of the 16 affected patients had been taking an antithrombotic drug before surgery. cSDH recurred in 52 cases (26%) in the group with and in 74 cases (38%) in the group without prior antithrombotic medication. In the univariate model, the following factors were associated with the occurrence of thrombotic complications: pre-existing neurological disease, pre-existing pulmonary disease, pre-existing malignancy, pre-existing cardiac disease/arrhythmia, preoperative paresis, Markwalder score (dichotomized), and the use of antithrombotic drugs. The parameter “cardiac pre-existing conditions/cardiac arrhythmias” was not included in the multivariable model because of its correlation with antithrombotic drug use (r = 0.465; r = 0.386); instead, the variable “antithrombotic medication” served as a surrogate parameter for cardiovascular disease. In the multivariable model, the use of antithrombotic drugs before the intervention and their interruption up to the time of the intervention or beyond it (until recovery) was a highly significant predictor of thromboembolic complications (odds ratio [OR]: 8.28; 95% confidence interval: [1.8; 38.12]; p = 0.007). Thus, the likelihood of a thrombotic complication was more than eight times higher in previously anticoagulated patients than in non-anticoagulated patients with cSDH. The variables of the regression analysis are summarized in eTable 3 (efigure 1).

Literature review

A selective search for publications in the PubMed and EMBASE databases over the period 1 January 2015 to 1 October 2020 was carried out with the following search strategy: (“chronic subdural hematoma” OR cSDH) AND (“anticoagulant withdrawal” OR “anticoagulant reversal” OR anticoagulation OR antithrombotic OR antiplatelet). The aim was to identify clinical studies (randomized controlled trials, cohort studies, case-control studies, and case series of more than 10 patients) of patients with cSDH focusing on the management of antithrombotics, specifically including documentation of thromboembolic events during the course. The inclusion criteria were as follows: clinical study (with at least 10 patients), patients over 18 years old, patient population with cSDH, with available information on antithrombotics as well as on thromboembolic complications and on outcomes at least six weeks postoperatively. After initial screening of all abstracts by two of the authors (HH, AH), the full text of all articles bearing a relation to antithrombotic medication was analyzed (HH, AH). The type of study, number of cases, period of patient enrollment, age and sex distribution, number of patients on antithrombotics (including specific drugs), management of antithrombotics, thromboembolic complications, study outcomes, and conclusions were extracted. In addition, the risk of bias was estimated on the Newcastle-Ottawa Scale (NOS) for non-randomized trials (HH, AH). The primary outcome parameter was the number of thromboembolic complications as a function of the mode of management of anticoagulation. Because of the heterogeneity of anticoagulation management across trials, a quantitative analysis of the primary outcome parameter was not possible. The results with respect to thromboembolic complications were presented separately, including case numbers, antithrombotic drug management, odds ratios, and 95% confidence intervals (95% CI); data were summarized narratively. The literature search was registered in advance in the PROSPERO database (CRD42021215982). The literature review was performed as recommended in the PRISMA statement (18). The database query was performed on 18 January 2021 (eFigure 2, eTable 4).

eTable 1. Markwalder classification (e2).

Markwalder score	Findings
0	neurologically intact
1	alert and oriented; mild symptoms such as headache; absent or mild neurological deficit
2	drowsy or disoriented with variable neurological deficit
3	stuporous but responding appropriately to noxious stimuli; severe focal signs such as hemiplegia
4	comatose with absent motor responses to painful stimuli; decerebrate or decorticate posturing

eTable 2. Distribution of variables (stated as total numbers and percentages, with means and standard deviations) for the overall patient collective, with and without antithrombotic medication (AT).

	Overall (n = 395)	No AT (n = 197)	AT (n = 198)	p-value*
Age	75 ± 12	72 ± 14	78 ± 9	< 0.0001
Sex (female)	145 (36.7%)	82 (41.6%)	63 (31.8%)	0.048
Markwalder score	1.6 ± 0.7	1.5 ± 0.7	1.7 ± 0.7	0.006
CHA2DS2-VASc score	3.4 ± 1.9	2.5 ± 1.6	4.3 ± 1.7	< 0.0001
GCS score	14.2 ± 2.1	14.3 ± 1.8	14.0 ± 2.3	0.185
Thromboembolic complications	20 (5.1%)	2 (1%)	18 (9.1%)	0.0003
Preoperative paresis	210 (53.7%)	98 (49.7%)	111 (56.1%)	0.264
Cardiac arrhythmias	95 (24.1%)	15 (7.6%)	80 (40.4%)	<0.001
Prior heart disease	138 (34.9%)	25 (12.7%)	113 (57.1%)	< 0.001
Prior lung disease	37 (9.4%)	18 (9.1%)	19 (9.6%)	1
Prior malignant disease	60 (15.2%)	32 (16.1%)	28 (14.1%)	0.674
Prior neurological disease	75 (19.1%)	45 (22.8%)	30 (15.2%)	0.072
Alcohol abuse	19 (4.8%)	13 (6.6%)	6 (3%)	0.157
Drug abuse	8 (2%)	3 (1.5%)	5 (2.5%)	0.724

*p-values for comparisons of the groups with and without thrombotic complications (t test, chi-squared test, Mann-Whitney U test); significant findings with respect to the primary endpoint are in boldface.

GCS, Glasgow Coma Scale.

eTable 3. Analysis of predictors in the multivariable model: statistical significance, odds ratios, and 95% confidence intervals.

	p-value	Odds ratio	95% CI
Age	0.898	1.004	[0.95; 1.06]
Sex	0.449	1.536	[0.51; 4.66]
Prior lung disease	0.201	0.435	[0.12; 1.56]
Prior malignant disease	0.134	0.410	[0.13; 1.32]
Prior neurological disease	0.245	3.425	[0.43; 27.30]
Markwalder score (dichotomized)	0.064	3.902	[0.92; 16.42]
Preoperative paresis	0.972	1.021	[0.30; 3.24]
Antithrombotic medication	0.007	8.282	[1.80; 38.12]

CI, confidence interval

Questions on the article in issue 12/2022: Chronic Subdural Hematoma – Antithrombotics and Thrombotic Complications.

The submission deadline is 24 March 2023.

Only one answer is possible per question. Please select the answer that is most appropriate.

Question 1

Where is the fluid collection of a chronic subdural hematoma located?

1. between the dura mater and the skull

2. between the dura mater and the arachnoid

3. in the subarachnoid space

4. between the pia mater and the gray matter

5. in the superior sagittal sinus

Question 2

According to a Finnish study, what was the incidence of chronic subdural hematoma among 80- to 89-year-olds in the period 2011–2015?

1. 18.2/100 000

2. 52.1/100 000

3. 80.4/100 000

4. 130.3/100 000

5. 241.5/100 000

Question 3

How should anticoagulation be managed in the presence of a chronic subdural hematoma?

1. There is no generally accepted regimen to date for managing antithrombotics in the presence of a chronic subdural hematoma.

2. Antithrombotic medication can continue to be given without change.

3. Antithrombotic medication should be discontinued and antagonized.

4. Further antithrombotic therapy should be solely carried out with a DOAC.

5. Anticoagulation should be intensified.

Question 4

What is the approximate recurrence rate of surgically treated chronic subdural hematomas?

1. 1–5%

2. 5–10%

3. 10–20%

4. 20–30%

5. 40–50%

Question 5

In the cohort study presented here, what was the sex distribution of patients with chronic subdural hematoma?

1. ca. 2/3 were women

2. ca. 2/3 were men

3. approximately half were women

4. ca. 90 % were women

5. ca. 90 % were men

Question 6

What nonspecific symptoms of chronic subdural hematoma are mentioned in the text?

1. headache, gait disturbance, cognitive impairment

2. dizziness, nausea, headache

3. visual perceptual disturbances, headache, nausea

4. hallucinations, headache, vomiting

5. tremor, nausea, stuttering

Question 7

In the cohort study presented here, how high was the risk of thrombotic complications in patients with chronic subdural hematoma (cSDH) who had received antithrombotic treatment prior to their diagnosis, compared to patients with cSDH who had not been under antithrombotics?

1. half as high

2. four times as high

3. one-fourth as high

4. eight times as high

5. one-eighth as high

Question 8

How was antithrombotic medication managed perioperatively in patients undergoing surgery for cSDH in most of the studies discussed in the article?

1. It was temporarily interrupted (paused).

2. It was still given, but at a lower dose.

3. It was continued unchanged

4. It was continued at a higher dose.

5. It was switched to ASA perioperatively.

Question 9

Which of the following may be a therapeutic option for patients with cSDH who have a vital indication for antithrombotic therapy with multiple drugs?

1. embolization of the anterior communicating artery

2. embolization of the middle cerebral artery

3. embolization of the posterior cerebral artery

4. embolization of the middle meningeal artery

5. embolization of the circle of Willis

Question 10

What type of imaging study usually leads to the diagnosis of a chronic subdural hematoma and is most commonly used for further follow-up?

1. magnetic resonance imaging

2. computed tomography with intravenous contrast

3. transcranial ultrasound

4. electroencephalography

5. noncontrast computed tomography